cover crops influence soil microorganisms and phytoextraction of copper from a moderately...

10
Cover crops inuence soil microorganisms and phytoextraction of copper from a moderately contaminated vineyard K.A. Mackie a, , H.P. Schmidt b , T. Müller c , E. Kandeler a a Institute of Soil Science and Land Evaluation, Soil Biology Section, University of Hohenheim, Emil-Wolff-Strasse 27, 70599 Stuttgart, Germany b Ithaka Institute, La Place 92, 1966 Ayent, Switzerland c Institute of Crop Science, University of Hohenheim, Fruwirthstrasse 20, 70599 Stuttgart, Germany HIGHLIGHTS This is one of the few studies used to measure Cu phytoextraction in situ in a vineyard. Variation of cover crop biomass determined Cu phytoextraction potential. Soil microorganisms showed Cu tolerance at moderate pollution levels (135 mg Cu kg 1 ). Nutrient resources and environmental factors regulated soil microorganisms. abstract article info Article history: Received 1 April 2014 Received in revised form 3 July 2014 Accepted 25 August 2014 Available online xxxx Editor: Charlotte Poschenrieder Keywords: PLFA Enzyme activity Microbial biomass PCA We investigated the ability of summer (Avena sativa [oat], Trifolium incarnatum [crimson clover], Chenopodium [goosefoot]) and winter (Vicia villosa [hairy vetch], Secale Cereale L. [Rye], Brassica napus L. partim [rape]) cover crops, including a mixed species treatment, to extract copper from an organic vineyard soil in situ and the microbial communities that may support it. Clover had the highest copper content (14.3 mg Cu kg 1 DM). However, it was the amount of total biomass production that determined which species was most effective at overall copper removal per hectare. The winter crop rye produced signicantly higher amounts of biomass (3532 kg DM ha 1 ) and, therefore, removed signicantly higher amounts of copper (14,920 mg Cu ha 1 ), despite less accumulation of copper in plant shoots. The maximum annual removal rate, a summation of best performing summer and winter crops, would be 0.033 kg Cu ha 1 y 1 . Due to this low annual extraction efcien- cy, which is less than the 6 kg Cu ha 1 y 1 permitted for application, phytoextraction cannot be recommended as a general method of copper extraction from vineyards. Copper concentration did not inuence aboveground or belowground properties, as indicated by sampling at two distances from the grapevine row with different soil copper concentrations. Soil microorganisms may have become tolerant to the copper levels at this site. Microbial biomass and soil enzyme activities (arylsulfatase and phosphatase) were instead driven by seasonal uxes of resource pools. Gram+ bacteria were associated with high soil moisture, while fungi seemed to be driven by extractable carbon, which was linked to high plant biomass. There was no microbial group associated with the increased phytoextraction of copper. Moreover, treatment did not inuence the abundance, activity or community structure of soil microorganisms. © 2014 Elsevier B.V. All rights reserved. 1. Introduction Due to the long history of application and continued use of copper containing fungicides in agriculture, copper (Cu) has accumulated with- in these topsoils (McBride et al., 1981; Mackie et al., 2012). Moderate levels of Cu have been shown to negatively affect macro-organisms, such as earthworms and plants, specically in biomass and seed set, as well as organic matter decomposition (Moolenaar, 1998; Paoletti et al., 1998; Brun et al., 2003; Hinojosa et al., 2010). It severely decreases the functional diversity of the soil microbial community, impairs specif- ic pathways of nutrient cycling and impacts soil fertility indicators at amounts as low as 140 mg Cu kg 1 (Kandeler et al., 1996; Fernández-Calviño et al., 2010; Hinojosa et al., 2010; Mackie et al., 2013). For these reasons, the European Union has set a limit on the amount of copper fungicide permitted for use in agriculture at 6 kg ha 1 y 1 (European Commission, 2007). However, as there are currently no viable alternatives in organic agriculture (Heibertshausen et al., 2006; La Torre et al., 2007) and as the potential for infection Science of the Total Environment 500501 (2014) 3443 Corresponding author at: Emil-Wolff-Straße 27, 70599 Stuttgart, Germany. Tel.: +49 711 459 22825. E-mail address: [email protected] (K.A. Mackie). http://dx.doi.org/10.1016/j.scitotenv.2014.08.091 0048-9697/© 2014 Elsevier B.V. All rights reserved. Contents lists available at ScienceDirect Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Upload: e

Post on 21-Feb-2017

212 views

Category:

Documents


0 download

TRANSCRIPT

Science of the Total Environment 500–501 (2014) 34–43

Contents lists available at ScienceDirect

Science of the Total Environment

j ourna l homepage: www.e lsev ie r .com/ locate /sc i totenv

Cover crops influence soil microorganisms and phytoextraction of copperfrom a moderately contaminated vineyard

K.A. Mackie a,⁎, H.P. Schmidt b, T. Müller c, E. Kandeler a

a Institute of Soil Science and Land Evaluation, Soil Biology Section, University of Hohenheim, Emil-Wolff-Strasse 27, 70599 Stuttgart, Germanyb Ithaka Institute, La Place 92, 1966 Ayent, Switzerlandc Institute of Crop Science, University of Hohenheim, Fruwirthstrasse 20, 70599 Stuttgart, Germany

H I G H L I G H T S

• This is one of the few studies used to measure Cu phytoextraction in situ in a vineyard.• Variation of cover crop biomass determined Cu phytoextraction potential.• Soil microorganisms showed Cu tolerance at moderate pollution levels (135 mg Cu kg−1).• Nutrient resources and environmental factors regulated soil microorganisms.

⁎ Corresponding author at: Emil-Wolff-Straße 27, 7059711 459 22825.

E-mail address: [email protected] (K.A. Ma

http://dx.doi.org/10.1016/j.scitotenv.2014.08.0910048-9697/© 2014 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 1 April 2014Received in revised form 3 July 2014Accepted 25 August 2014Available online xxxx

Editor: Charlotte Poschenrieder

Keywords:PLFAEnzyme activityMicrobial biomassPCA

We investigated the ability of summer (Avena sativa [oat], Trifolium incarnatum [crimson clover], Chenopodium[goosefoot]) and winter (Vicia villosa [hairy vetch], Secale Cereale L. [Rye], Brassica napus L. partim [rape]) covercrops, including a mixed species treatment, to extract copper from an organic vineyard soil in situ and themicrobial communities that may support it. Clover had the highest copper content (14.3 mg Cu kg−1 DM).However, it was the amount of total biomass production that determined which species was most effective atoverall copper removal per hectare. The winter crop rye produced significantly higher amounts of biomass(3532 kg DM ha−1) and, therefore, removed significantly higher amounts of copper (14,920 mg Cu ha−1),despite less accumulation of copper in plant shoots. The maximum annual removal rate, a summation of bestperforming summer andwinter crops,would be 0.033 kg Cu ha−1 y−1. Due to this low annual extraction efficien-cy,which is less than the 6 kg Cu ha−1 y−1 permitted for application, phytoextraction cannot be recommended asa general method of copper extraction from vineyards. Copper concentration did not influence aboveground orbelowground properties, as indicated by sampling at two distances from the grapevine row with different soilcopper concentrations. Soil microorganismsmay have become tolerant to the copper levels at this site. Microbialbiomass and soil enzyme activities (arylsulfatase and phosphatase) were instead driven by seasonal fluxes ofresource pools. Gram+ bacteria were associated with high soil moisture, while fungi seemed to be driven byextractable carbon, which was linked to high plant biomass. There was no microbial group associated with theincreased phytoextraction of copper. Moreover, treatment did not influence the abundance, activity orcommunity structure of soil microorganisms.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

Due to the long history of application and continued use of coppercontaining fungicides in agriculture, copper (Cu) has accumulatedwith-in these topsoils (McBride et al., 1981; Mackie et al., 2012). Moderatelevels of Cu have been shown to negatively affect macro-organisms,such as earthworms and plants, specifically in biomass and seed set, as

9 Stuttgart, Germany. Tel.: +49

ckie).

well as organic matter decomposition (Moolenaar, 1998; Paolettiet al., 1998; Brun et al., 2003; Hinojosa et al., 2010). It severely decreasesthe functional diversity of the soil microbial community, impairs specif-ic pathways of nutrient cycling and impacts soil fertility indicators atamounts as low as 140 mg Cu kg−1 (Kandeler et al., 1996;Fernández-Calviño et al., 2010; Hinojosa et al., 2010; Mackie et al.,2013). For these reasons, the European Union has set a limit onthe amount of copper fungicide permitted for use in agriculture at6 kg ha−1 y−1 (European Commission, 2007). However, as there arecurrently no viable alternatives in organic agriculture (Heibertshausenet al., 2006; La Torre et al., 2007) and as the potential for infection

35K.A. Mackie et al. / Science of the Total Environment 500–501 (2014) 34–43

from plant pathogens increases with climate change (Salinari et al.,2006), Cu fungicides have not been prohibited and may even increasein the future.

In response to these negative effects, one possible solution is Cu re-moval through in situ accumulation by plants. Phytoextraction is theuse of (hyper) accumulator plants to remove metals/metalloids fromthe environment by taking them up into their shoots and subsequentlyremoving them from the contaminated area (Wenzel, 2009). It is a lowcost, environmentally sensitive method, which displaces Cu from theenvironment, but does not require full soil removal impractical in pe-rennial agriculture and/or large tracts of land (Gómez-Sagasti et al.,2012; Meier et al., 2012a).

Particular microorganisms prefer specific plants and plant speciessupport and encourage associated microorganisms (Terry and Bañuelos,2000; Wardle et al., 2004; Castaldi et al., 2009; Narula et al., 2009;Epelde et al., 2010; Haferburg and Kothe, 2010). Themost recent mecha-nism for enhancing phytoextraction is inoculating the soil with bacteriaproducing siderophores, which assist in chelating Cu, suggesting that mi-croorganisms may play a significant role in successful phytoextraction(Haferburg and Kothe, 2010; Rajkumar et al., 2010). Phytoextraction,with and without inoculated microbial assistance, has been successfullyinvestigated in laboratories and greenhouses (Poschenrieder et al.,2001; Brun et al., 2003; Kos and Leštan, 2004; Song et al., 2004; Chenet al., 2006; Meier et al., 2012b; Ma et al., 2009; Zeremski-Škorić et al.,2010; Andreazza et al., 2011). However, phytoextraction of Cu has sel-dom been monitored in the field (Poschenrieder et al., 2001; Clementeet al., 2005; Brej and Fabiszewski, 2006).

The aims of this study were to investigate the in situ relationship be-tween microorganisms and plants within Cu contaminated topsoil, andidentify the practicability of phytoextraction and monitor ecosystem ser-vices, such as soil health and nutrientmineralization, usingmicrobial bio-mass, enzyme activity and phospholipid fatty acids (PLFAs) (Epelde et al.,2014). Enzyme activities are consistent biological indicators of heavymetal pollution and PLFA patterns have been seen to change quicklywith changing soil metal concentration in as little as two weeks(Frostegård et al., 1996; Hinojosa et al., 2010; Ge and Zhang, 2011). Addi-tionally, PLFAs identify microbial groups, which may indicate whethersuch groups naturally support increasedphytoextraction of specific plantsin situ. This project focusedparticularly onvineyards, a representative sys-tem of fruit productionwhere Cu is most often applied. In vineyards withsufficient water, i.e. central Europe, cover crops grown between the vinerows have been observed to increase desirable properties in soil andvine performance (Morlat and Jacquet, 2003; Guerra and Steenwerth,2012). Therefore, phytoextraction has the potential to improve grape pro-duction and soil fertility in vineyards, while removing Cu from the topsoil.We investigatedwhether (i) the efficiency of phytoextraction depends onplant species, plant community composition, distance from vine row, andgrowing season, (ii) if effective phytoextraction is associated with a mi-crobial community structure, and (iii) if diverse plant communities willmitigate the negative influence of Cu on soil microorganisms.

The plants chosen within the present study were a mixture of Cuadapted plant species with high biomass production known from labo-ratory research as well as common vineyard cover crop species in cen-tral Europe not yet researched for Cu removal potential (Poschenriederet al., 2001; Kos and Leštan, 2004; Andreazza et al., 2010; Haferburg andKothe, 2010).Moreover, diverse plant systems, in comparison tomono-culture systems, have been seen to reduce the impact of pollution by in-creasingmicrobial diversity and activity (Yang et al., 2007). Therefore, atreatment consisting of a mixture of plant species has also been added.

2. Materials & methods

2.1. Study site and experimental setup

This field experiment was designed specifically to understand thein situ potential of cover crop plant species to accumulate Cu and was

established at the Ithaka Institute in Canton Wallis, Switzerland(46°16′N, 7°24′E) in the spring of 2012. The study site is a vineyardplanted with Pinot noir (Vitis vinifera L.) with a southeastern exposureat an elevation of 760–780m a.s.l. The site has a mean annual precipita-tion of 550 mm and an average temperature of 11.4 °C. The soil is pre-dominately calcaric Leptosol with a bulk density of 1.34 g cm−3 and47% gravel (N2 mm). It has a pHCaCl2 of 7.5, CT of 37 g kg−1, NT of4.1 g kg−1, and a total microbial biomass of 694 μg Cmic g−1 soil. Thevines are planted at a distance of 1 m and a row width of 3 m with anadapted Mosel arch training system. At their maximum, vine plantheight from July until October averages 2.50m. The Ithaka Instituteman-ages the vineyard organically; however, for this field experiment the useof copper fungicides was excluded during the trial period so that thesampleswould not be superficially contaminated. Plant and compost de-rived solutions (100 L compost tea ha−1, 100 g NU-Film ha−1,2 kg stinging nettle ha−1, 1 kg horsetail ha−1, and 100 g sage ha−1)and sodiumbicarbonatewere sprayed to stimulate the plants natural de-fenses and protect against Oidium and Peronospora, respectively. The ini-tial total Cu in soilwas 135mgCuT kg−1 soil (95.9 kg CuT ha−1),while theexchangeable Cu fraction was initially 48.6 mg CuDTPA kg−1 soil(34.5 kg CuDTPA ha−1). Soil mass was calculated using soil depth, bulkdensity and fraction of coarse material (N2 mm) in order to calculatekg soil Cu ha−1. The site was superficially tilled (8 cm) before seedingin April 2012.

We designed four summer (2012) treatments followed succes-sively by four winter (2012/2013) treatments. The summer cropswere seeded in April 2012 and harvested in August 2012; a) Avenasativa (oat) 12 g m−2, b) Trifolium incarnatum (clover) 3 g m−2,c) Reseda luteola (Reseda) 0.3 g m−2, and d) a mixture of treatments1, 2 and 3 (summer mix) at 4, 1 and 0.1 g m−2, respectively. Unfortu-nately, although a drought tolerant species ofResedawas chosen, the spe-cies did not germinate well in all plots and Chenopodium album L.,Chenopodiumhybridum L. andChenopodium fielfolium Sm. (Chenopodium)spontaneously took over. Therefore, Chenopodium was harvested andsampled in both treatments c and d. The winter crops were seeded inSeptember 2012 and harvested in May 2013; a) Vicia villosa (Hairyvetch) 10 g m−2, b) Secale Cereale L. (Rye) 20 g m−2, c) Brassica napusL. partim (Rape) 0.6 g m−2, and d) a mixture of treatments 1, 2 and3 at 3.3, 6.7 and 0.2 g m−2, respectively (winter mix). The treatmentswere set up in a random block design, where each plot had an area of36 m2 and spanned three vine rows. The treatments were replicatedfive times. Plant and soil samples were taken over three samplingdates: June 2012 (normal mowing date for summer cover crop), August2012 (extended mowing date and plant senescence), and May 2013(normalmowing date forwinter cover crop). As there is a buffer time be-tween the summer crop harvest andwinter crop seeding, two dateswerechosen to assesswhether the extended time could achieve higher extrac-tion rates. As summer crop seeding has to be done within a specific timewindow, there was no possibility to extend the winter crop harvest andonly one date was chosen. Samples were taken on either side of themid-dle vine row, reducing edge effects, at both 70 cm and 120 cm away fromthe vine row in order to represent high Cu areas and high plant biomassareas, respectively (S.1).

2.2. Plant sampling and analyses

Plant samples were taken directly above and surrounding the soilcorer with the guidance of a plastic ring ( = 30.5 cm) and were re-moved prior to soil sampling. All living above ground biomass was cutand placed directly into paper bags. The samples were placed in a60 °C oven for at least 48 h after which they were weighed for drymass (DM).

Dried plant material was finely milled (SM1 Schneidemühle, RetschGmbH, Germany) and underwent microwave digestion in nitric acidfollowing the method of VDLUFA (2011a). To determine the Cu contentof the plant biomass the method of VDLUFA (2011b) was used and

36 K.A. Mackie et al. / Science of the Total Environment 500–501 (2014) 34–43

analyzed with inductively coupled plasma optical emission spectrome-try (ICP-OES, PerkinElmer, USA).

2.3. Soil sampling and analyses

Four soil cores ( =5.5 cm), at a depth of 10 cm, from each plot, weretaken to generate a plot bulk sample. The cores were taken following arandom sample design and in order to exclude surface plant residues,the top one-centimeter was removed from each soil core. Sampleswere kept in a cooler and placed in a 4 °C refrigerator after sampling.The soil was sieved at 2 mm and homogenized. Finally, the sampleswere stored in a −20 °C freezer for further analyses. Samples for Cuanalyses were air-dried after sieving.

Total copper (CuT) was extracted by Aqua regia (HNO3 + HCl) ex-tractant (DIN ISO, 11466, 1995). As an indicator of bioavailable copper,the DTPA extractable copper fraction (CuDTPA) was analyzed usingdiethylenetriamine pentaacetic acid extractant (CaCl2 + DTPA)(VDLUFA, 2011c). DTPA is an effective extractant for neutral and slightlyalkaline soils and is used for extracting ligand-bound metals within thesoil (Brun et al., 1998; Ettler et al., 2007). All analyses of extractantswere done with an atomic absorption spectrophotometer. pH was de-termined with a glass electrode in a suspension of 4 g soil in 10 ml0.01MCaCl2. Soilwater contentwasdetermined gravimetrically by dry-ing samples at 105 °C for 24 h. Samples dried at 105 °C were finelyground and weighed to 0.060 g to be analyzed for total carbon (CT)and nitrogen (NT) using dry combustion (element analyzer, ElementarAnalysensysteme GmbH, Germany).

The chloroform fumigation extraction (CFE) method of Vance et al.(1987)was used to estimate andmeasuremicrobial (Cmic, Nmic) and ex-tractable organic carbon (EOC) and nitrogen (ETN), using a fresh soilweight of 10 g. Chloroform fumigated and non-fumigated sampleswere extracted with 40 ml 0.5 M K2SO4 on a shaker (250 U min−1) for30 min. After being centrifuged at 4400 g for 30 min a 1:4 dilution ofthe supernatant was analyzed using a TOC-TNb Analyzer Multi-N/C2100S (Analytik Jena, Germany). 1MHCl was added to the sample dilu-tions before measurement to remove small amounts of inorganic C.Since only visible roots were removed prior to fumigation and extrac-tion, it cannot fully be excluded that chloroform labile C and N werecontaminated by C and N derived from fine roots remaining in the soilsample (Mueller et al., 1992). The estimation of Cmic and Nmic used kecfactors of 0.45 and 0.54, respectively (Joergensen, 1996).

The following enzyme analyses were measured using a spectropho-tometer (UV-1601 Spectrophotometer, Shimadzu, Germany).Arylsulfatase, measured according to Schinner et al. (1996), was ana-lyzed to assess the mineralization of organic sulfur compounds. Onegram of soil, as suggested for marginal soils, was mixed with acetatebuffer and 4-nitrophenylsulfate and incubated for 1 h at 37 °C. Sampleswere then mixed with water and filtered and sodium hydroxide wasadded before analysis. Phosphatase is a component of soil enzymesthat removes phosphate from its organic substrate. Whereas soil micro-organisms exclusively produce alkaline phosphatases, either soil micro-organisms or plants can produce acid phosphatases. In this study thealkaline phosphatase was measured using the method of Schinneret al. (1996). Soil, 0.3 g, was mixed with borate buffer andphenylphosphate-disodium salt and incubated for 3 h at 37 °C. Sampleswere filtered and mixed with 2,6-dibromchinon-chloromid, the colorwas allowed to develop for 30 min before analyzing.

Four grams of soil was taken for lipid extraction and fractionationfollowing the alkaline methylation method according to Frostegardet al. (1991). The resulting phospholipid fatty acid (PLFA)methyl ethers(MEs) were dissolved in Isooctan and measured by gas chromatograph(Auto System XL, PerkinElmer, USA) using an HP-5 capillary column, aflame ionization detector and helium as the carrier gas. FAMEs wereidentified using their retention time based on fatty- and bacterial-acidmethylester-mix. Quantificationwas calculatedwith the use of an inter-nal FAME standard, which had been added before methanolysis.

Nomenclature and division of PLFAs into bacteria and fungi werebased on Kandeler et al. (2008), Frostegård and Bååth (1996), andZelles (1999). Gram+ PLFAs (PLFAgram+) were represented by i15:0,a15:0, i16:0, and i17:0. Gram− (PLFAgram−) were represented bycy17:0 and cy19:0. Total bacterial PLFAs (PLFAbacteria) were representedby the sum of PLFAgram−, PLFAgram+ and 16:1ω7. Fungal PLFAs(PLFAfungal) were represented by 18:2ω6.

2.4. Statistics

The R statistical programwas used for all statistical analyses (R CoreTeam, 2013). Initial sampling in April 2012was tested for significant dif-ferences using a one-way ANOVA with treatment as a fixed factor. Astreatment was never found to be significant, the average mean wastaken of each variable and given as baseline soil data (Section 2.1; CT,NT, Cmic, CuT, CuDTPA) and not taken as a co-variable in the statisticalanalysis. To analyze differences between a normal mowing date andan extended date during the summer crop, June 2012 and August2012were tested using a three-way ANOVA, where treatment, distanceand date were fixed factors; a time series analysis was used. The samestatistical analyses were used in order to compare overall differencesbetween summer and winter crops, June 2012 and May 2013, withthe exception that date was now considered as crop season. June 2012was selected, instead of August 2012, because it had green abovegroundvegetation and was the normal mowing date, both of which were alsotrue of May 2013. Block and column were used as fixed factors inevery case.When significant differenceswere observed, Tukey's HonestSignificant Distance test was performed so that differences could bespecified. Significance was tested for p b 0.05 in all cases.

In order to determine how microbial communities were affected bydate, treatment and distance from the vine row, multivariate statisticalanalysis (MANOVA) was conducted. Each individual PLFA (see aboveSection 2.6) was used for the MANOVA. When significant results wereobtained, all individual PLFAs were normalized and principal compo-nents analysis (PCA) was performed. The PLFA loadings for the firsttwo axeswere examined to determinewhich PLFAs weremost stronglyassociated with each axis. The scores of the first two PCs were then cor-related with soil chemical properties and plant properties measured tosee which were significantly correlated with each PC. Significance wastested for p b 0.05 in all cases.

3. Results

3.1. Aboveground vegetation and copper uptake

3.1.1. Summer crop

3.1.1.1. Plant copper content. Plant shoots contained varying concentra-tions of Cu, ranging in mean from 8.4 to 14.3 mg Cu kg−1 DM betweenJune 2012 and August 2012 (Fig. 1a). Treatment was highly significant(Table 1). Clover shoots had the highest Cu content (14.3 mg Cu kg−1

DM), which was significantly more than summer mix, oat andChenopodium. Summer mix shoots had the second highest Cu content(11.8 mg Cu kg−1 DM), significantly more than Chenopodium. Plant Cucontent increased significantly from June 2012 to August 2012.

3.1.1.2. Plant shoot biomass and plant copper removal. Plant shoot bio-mass was only significantly influenced by the interaction between dis-tance from the vine and plant treatment (Table 1). The treatments hadbiomass ranging from 311–1476 kg DM ha−1, where oat—120 cm(1476 kg DM ha−1) was significantly higher than all other treatmentsand distances (Fig. 1b). There was a negative correlation betweenplant biomass and plant Cu content (r = −0.59, p b 0.001).

Plant Cu removal was significantly influenced by the interaction be-tween plant treatment and distance from the vine row (Table 1). Oat at120 cm (13,620 mg Cu ha−1) was found to be significantly better at Cu

Fig. 1. Plant copper content (a), plant biomass (b) and plant copper removal (c) during the summer crops (June and August 2012) andwinter crop (May 2013) at a distance of 70 cm and120 cm from the vine row. Where O is oat, CC is clover, CH is Chenopodium, SM is summer mix, HV is hairy vetch, RY is rye, RA is rape, and WM is winter mix. Mean ± SE is shown.

37K.A. Mackie et al. / Science of the Total Environment 500–501 (2014) 34–43

removal when compared to all other treatments and distances (3357–4983 mg Cu ha−1), with the exception of clover—120 cm(8477 mg Cu ha−1), which was not significantly different from anytreatment (Fig. 1c). Cu removal had a significant positive correlationwith plant shoot biomass (r = 0.84, p b 0.001).

3.1.2. Winter crop

3.1.2.1. Plant copper content. Plant shoot Cu content was significantly in-fluenced by plant treatment and crop season (Table 2). The summercrop treatments were generally significantly higher than those of thewinter crop treatments, regardless of distance from the vine row.Hairy vetch Cu content was significantly higher than other winter

Table 1Statistical significance of the three-way ANOVA. Treatment, distance and date, including theirTreatment ∗ date, distance ∗ date and treatment ∗ distance ∗ date were not significant in any c

Treatment(3 d.f.)

Distance(1 d.f.)

F-value p-Value F-value p-Value

Plant Cu content 6.90 0.0016 n.s. n.s.Plant shoot biomass 6.45 0.0025 11.39 0.0026Plant Cu removal 4.70 0.0106 7.40 0.0122Cmic n.s. n.s. 6.22 0.0182Arylsulfatase n.s. n.s. 27.41 b0.0001Phosphatase n.s. n.s. 42.80 b0.0001EOC n.s. n.s. 7.47 0.0103ETN n.s. n.s. n.s. n.s.CuT n.s. n.s. 52.65 b0.0001CuDTPA n.s. n.s. 64.11 b0.0001Nmic n.s. n.s. 6.03 0.0199PLFAbacteria n.s. n.s. 15.07 0.0006PLFAfungal n.s. n.s. 10.33 0.0031PLFAgram + n.s. n.s. 22.46 0.0001PLFAgram− n.s. n.s. 10.04 0.0034Soil moisture n.s. n.s. 11.53 0.0019

treatments, but it was significantly lower than the highest summercrops (Fig. 1a).

3.1.2.2. Plant shoot biomass and plant copper removal. The interaction be-tween plant treatment and crop season significantly influenced plantshoot biomass (Table 2). Rye had significantly higher biomass than allother cover crops. Hairy vetch was also significantly higher than thesummer crops (Fig. 1b). Distance was significant for both crop seasons;treatments grew more at 120 cm than at 70 cm.

When comparing rates of plant Cu removal, treatment, cropseason and distance were significant (Table 2). Winter crops(10,930 mg Cu ha−1) removed significantly more Cu than summercrops (6247 mg Cu ha−1) (Fig. 1c). Overall, there was also signifi-cantly higher Cu removal at 120 cm than at 70 cm from the

interactions for each variable, during the summer crop (June and August 2012) is shown.ase. Significance is shown only for a p b 0.05 and n.s. is not significant.

Crop season(1 d.f.)

Trt ∗ crop season(3 d.f.)

Dist ∗ crop season(1 d.f.)

F-value p-Value F-value p-Value F-value p-Value

316.62 b0.0001 29.91 b0.0001 n.s. n.s.64.69 b0.0001 7.34 0.0009 n.s. n.s.22.15 b0.0001 n.s. n.s. n.s. n.s.13.77 0.0006 n.s. n.s. n.s. n.s.33.55 b0.0001 n.s. n.s. n.s. n.s.n.s. n.s. n.s. n.s. n.s. n.s.74.56 b0.0001 n.s. n.s. 10.84 0.002214.10 0.0006 n.s. n.s. n.s. n.s.n.s. n.s. n.s. n.s. n.s. n.s.n.s. n.s. n.s. n.s. n.s. n.s.19.10 0.0001 n.s. n.s. n.s. n.s.14.77 0.0005 n.s. n.s. n.s. n.s.40.29 b0.0001 n.s. n.s. 6.04 0.01866.29 0.0171 3.31 0.0315 n.s. n.s.n.s. n.s. n.s. n.s. n.s. n.s.537.66 b0.0001 n.s. n.s. n.s. n.s.

Table 2Statistical significance of the three-way ANOVA. Treatment, distance and crop season, including their interactions for each variable, between the summer crop (June 2012) and thewintercrop (May 2013) are shown. Treatment ∗ distance and treatment ∗ distance ∗ crop season were not significant in any case. Significance is shown only for a p b 0.05 and n.s. is notsignificant.

Treatment(3 d.f.)

Distance(1 d.f.)

Crop season(1 d.f.)

Trt ∗ crop season(3 d.f.)

Dist ∗ crop season(1 d.f.)

F-value p-Value F-value p-Value F-value p-Value F-value p-Value F-value p-Value

Plant Cu content 6.90 0.0016 n.s. n.s. 316.62 b0.0001 29.91 b0.0001 n.s. n.s.Plant shoot biomass 6.45 0.0025 11.39 0.0026 64.69 b0.0001 7.34 0.0009 n.s. n.s.Plant Cu removal 4.70 0.0106 7.40 0.0122 22.15 b0.0001 n.s. n.s. n.s. n.s.Cmic n.s. n.s. 6.22 0.0182 13.77 0.0006 n.s. n.s. n.s. n.s.Arylsulfatase n.s. n.s. 27.41 b0.0001 33.55 b0.0001 n.s. n.s. n.s. n.s.Phosphatase n.s. n.s. 42.80 b0.0001 n.s. n.s. n.s. n.s. n.s. n.s.EOC n.s. n.s. 7.47 0.0103 74.56 b0.0001 n.s. n.s. 10.84 0.0022ETN n.s. n.s. n.s. n.s. 14.10 0.0006 n.s. n.s. n.s. n.s.CuT n.s. n.s. 52.65 b0.0001 n.s. n.s. n.s. n.s. n.s. n.s.CuDTPA n.s. n.s. 64.11 b0.0001 n.s. n.s. n.s. n.s. n.s. n.s.Nmic n.s. n.s. 6.03 0.0199 19.10 0.0001 n.s. n.s. n.s. n.s.PLFAbacteria n.s. n.s. 15.07 0.0006 14.77 0.0005 n.s. n.s. n.s. n.s.PLFAfungal n.s. n.s. 10.33 0.0031 40.29 b0.0001 n.s. n.s. 6.04 0.0186PLFAgram + n.s. n.s. 22.46 0.0001 6.29 0.0171 3.31 0.0315 n.s. n.s.PLFAgram− n.s. n.s. 10.04 0.0034 n.s. n.s. n.s. n.s. n.s. n.s.Soil moisture n.s. n.s. 11.53 0.0019 537.66 b0.0001 n.s. n.s. n.s. n.s.

38 K.A. Mackie et al. / Science of the Total Environment 500–501 (2014) 34–43

vine rows. The average means of summer mix and winter mix(5125 mg Cu ha−1) were significantly lower than the means of oatand hairy vetch (11,018 mg Cu ha−1) and clover and rye(11,039 mg Cu ha−1).

3.2. Soil chemical properties

3.2.1. Summer cropCT, NT, and ETN were significantly higher at 70 cm than at 120 cm

from the vine row over the summer crop season (June and August2012) (Tables 1, 3). EOC was not significantly influenced by any factor.Carbon and nitrogen were not correlated to any aboveground plantcharacteristics.

CuT and CuDTPA were significantly influenced by distance from thevine row (Tables 1, 3). At 70 cm Cu concentration was higher thanthat at 120 cm from the vine row. CuT and CuDTPAwere significantly pos-itively correlated to each other (r = 0.91, p b 0.001). CuT was signifi-cantly positively correlated with CT and NT (r = 0.75 and r = 0.69,respectively, p b 0.001), as was CuDTPA (r = 0.72 and r = 0.73, respec-tively, p b 0.001). CuDTPA was also significantly correlated to EON(r = 0.51, p b 0.001).

3.2.2. Winter cropCT and NT were significantly higher at 70 cm than at 120 cm

(Tables 2, 3). EOC, however, was influenced by the interaction betweencrop season and distance from the vine row. In May 2013, EOC wassignificantly higher at 70 cm (221 μg EOC g−1) than at 120 cm(166 μg EOC g−1). EOC was also significantly higher in May 2013 thanin June 2012. ETN significantly differed by both distance and crop sea-son. ETN content was significantly higher at 70 cm (29.4 μg ETN g−1)than at 120 cm (21.1 μg ETN g−1) and higher in May 2013 than inJune 2012.

CuT and CuDTPA were only significantly influenced by distance fromthe vine row, where values at 70 cm were higher than those at120 cm (Tables 2, 3).

3.3. Soil microbial abundance and function

3.3.1. Summer cropIn June and August 2012, microbial abundance only differed by date

(Table 1). Cmic and Nmic decreased significantly from June 2012 to Au-gust 2012 (Fig. 2a, Table 3). Cmic and Nmic were significantly correlated

to both EOC and ETN (r= 0.69, r= 0.67 and r= 0.60, r= 0.60, respec-tively, p b 0.001).

Microbial enzyme activities of arylsulfatase and phosphatase weresignificantly influenced by distance from the vine row; 70 cm was sig-nificantly higher than 120 cm from the vine row (Table 1, Fig. 2b andc). Arylsulfatase was significantly correlated to Cmic (r = 0.54,p b 0.001) as well as EOC (r = 0.54, p b 0.001).

3.3.2. Winter cropMicrobial abundance increased significantly from June 2012 to May

2013 (Fig. 2a, Table 3). It was also significantly influenced by distance(Table 2). Cmic and Nmic were significantly higher at 70 cm than at120 cm.

Arylsulfatase activity was also significantly higher at 70 cm than at120 cm (Fig. 2b, Table 2). Activity increased significantly from June2012 to May 2013. Phosphatase activity did not increase in May 2013and 120 cm remained significantly lower than 70 cm (Fig. 2c).

3.4. Microbial community

3.4.1. Summer cropBacterial PLFAs were significantly influenced by distance from the

vine row (Table 1). PLFAbacteria, PLFAgram+ and PLFAgram− were allhigher at 70 cm than at 120 cm. PLFAbacteria and PLFAgram+ also in-creased from June 2012 to August 2012 (Table 3). PLFAbacteria was posi-tively correlated to Cmic (r = 0.51, p b 0.001), CT (r = 0.58, p b 0.001),and NT (r= 0.55, p b 0.001). PLFAgram+was not correlated to any othervariable. PLFAgram−, on the other hand, was positively correlated to CT(r = 0.61, p b 0.001) and NT (r = 0.54, p b 0.001). PLFAfungal was theonly fraction not influenced by distance from the vine and wassignificantly higher in June 2012 than in August 2012. PLFAfungal wassignificantly positively correlated to Cmic (r = 0.71, p b 0.001), EOC(r = 0.69, p b 0.001) and ETN (r = 0.54, p b 0.001).

3.4.2. Winter cropAll measured PLFAs were significantly higher in May 2013 than in

June 2012, with the exception of Gram-bacteria (Table 3). In addition,distance from the vine row was highly significant, where bacterialPLFAs were higher at 70 cm than at 120 cm (Table 2). The interactionbetween treatment and crop season was significant for PLFAgram+,where rape in May 2013 had significantly higher abundance than oatin June 2012. For PLFAfungal there was a significant interaction betweendistance and crop season, where abundance was higher at 70 cm from

Table 3Carbon, nitrogen, extractable organic carbon (EOC), extractable total nitrogen (ETN), total soil copper, DTPA exchangeable copper,microbial nitrogen, and bacterial and fungal PLFAs.Meanconcentrations ± SE.

June 2012 August 2012 May 2013

Treatment 70 cm fromvine row

120 cm fromvine row

70 cm fromvine row

120 cm fromvine row

Treatment 70 cm fromvine row

120 cm fromvine row

C % Oat 5.0 ± 0.4 3.8 ± 0.3 6.0 ± 0.4 4.7 ± 0.4 Hairy vetch 4.7 ± 0.8 3.8 ± 0.2Clover 4.6 ± 0.4 4.0 ± 0.1 5.8 ± 0.4 5.2 ± 0.3 Rye 4.6 ± 0.4 3.7 ± 0.3Chenopodium 4.7 ± 0.6 3.9 ± 0.3 5.4 ± 0.3 4.3 ± 0.5 Rape 5.0 ± 0.3 4.0 ± 0.3Summer mix 4.5 ± 0.3 3.7 ± 0.3 5.6 ± 0.3 5.2 ± 0.4 Winter mix 4.8 ± 0.3 3.5 ± 0.1

N % Oat 0.51 ± 0.1 0.39 ± 0 0.50 ± 0 0.40 ± 0 Hairy vetch 0.45 ± 0.1 0.37 ± 0Clover 0.46 ± 0 0.41 ± 0 0.52 ± 0 0.50 ± 0 Rye 0.46 ± 0 0.38 ± 0Chenopodium 0.46 ± 0 0.41 ± 0 0.48 ± 0 0.37 ± 0.1 Rape 0.47 ± 0 0.40 ± 0Summer mix 0.46 ± 0 0.39 ± 0 0.50 ± 0 0.47 ± 0 Winter mix 0.46 ± 0 0.37 ± 0

EOC μg g−1 Oat 132 ± 29 121 ± 16 133 ± 30 115 ± 14 Hairy vetch 228 ± 31 168 ± 27Clover 107 ± 6.2 133 ± 13 130 ± 25 102 ± 14 Rye 200 ± 20 159 ± 13Chenopodium 125 ± 27 118 ± 20 121 ± 32 102 ± 11 Rape 243 ± 26 178 ± 14Summer mix 156 ± 13 131 ± 15 141 ± 31 98.7 ± 27 Winter mix 211 ± 15 160 ± 6

ETN μg g−1 Oat 19.8 ± 6.3 19.2 ± 3.6 30.6 ± 11 17.5 ± 3 Hairy vetch 38.0 ± 5.9 23.7 ± 4.3Clover 18.5 ± 2.8 18.0 ± 1.6 28.0 ± 4.2 14.9 ± 2.7 Rye 31.3 ± 4.6 21.5 ± 2.4Chenopodium 23.1 ± 6.5 18.4 ± 5 22.0 ± 3.6 14.3 ± 2.3 Rape 38.5 ± 3.7 25.4 ± 2.7Summer mix 33.6 ± 5.2 21.9 ± 5.1 26.1 ± 2.4 19.9 ± 1.8 Winter mix 32.5 ± 2.4 22.4 ± 1.2

CuT mg kg−1 Oat 150 ± 13 116 ± 9.7 179 ± 12 136 ± 13 Hairy vetch 162 ± 15 122 ± 4.9Clover 137 ± 14 115 ± 8.2 135 ± 45 132 ± 8.7 Rye 140 ± 9.6 115 ± 8.7Chenopodium 135 ± 12 108 ± 9.1 146 ± 13 127 ± 11 Rape 151 ± 9.7 120 ± 11Summer mix 140 ± 8.2 104 ± 7.6 150 ± 6.5 123 ± 5.9 Winter mix 131 ± 5.8 105 ± 6.6

CuDTPA mg kg−1 Oat 58.6 ± 5.3 43.6 ± 5.7 64.8 ± 4.9 43.0 ± 4.1 Hairy vetch 59.1 ± 7.2 38.4 ± 3.3Clover 51.8 ± 6.7 39.2 ± 2.1 54.7 ± 7.4 43.1 ± 4.1 Rye 50.3 ± 3.8 36.0 ± 3.1Chenopodium 49.1 ± 6.5 34.9 ± 3 53.7 ± 5.6 40.8 ± 4.3 Rape 51.4 ± 2.3 36.0 ± 3.9Summer mix 54.5 ± 4.1 36.8 ± 3.1 55.5 ± 2.7 42.3 ± 2.8 Winter mix 47.3 ± 1.0 34.9 ± 2.4

Nmic μg g−1 Oat 99.1 ± 11 163 ± 39 151 ± 26 135 ± 16 Hairy vetch 233 ± 25 159 ± 20Clover 146 ± 32 128 ± 19 113 ± 18 114 ± 14 Rye 200 ± 9.7 169 ± 14Chenopodium 158 ± 32 142 ± 38 124 ± 20 95.7 ± 7.9 Rape 214 ± 10 184 ± 16Summer mix 200 ± 18 139 ± 15 145 ± 37 94.9 ± 15 Winter mix 223 ± 14 175 ± 4.9

PLFAbacterial nmol g−1 DM Oat 51.7 ± 3.0 44.7 ± 4.3 64.8 ± 3.6 51.3 ± 4.0 Hairy vetch 70.4 ± 8.1 61.1 ± 7.9Clover 55.8 ± 3.5 46.1 ± 3.6 60.3 ± 2.6 58.9 ± 4.5 Rye 60.0 ± 3.5 49.0 ± 3.2Chenopodium 54.1 ± 6.9 62.8 ± 19 61.8 ± 5.8 52.8 ± 3.8 Rape 74.2 ± 4.8 58.8 ± 4.0Summer mix 56.7 ± 1.8 56.0 ± 6.8 61.8 ± 1.9 55.2 ± 3.4 Winter mix 63.3 ± 3.6 48.6 ± 1.2

PLFAgram+ nmol g−1 DM Oat 32.2 ± 1.9 27.5 ± 2.7 40.6 ± 2.1 32.3 ± 2.3 Hairy vetch 40.9 ± 4.6 39.5 ± 8.3Clover 35.2 ± 2.3 28.5 ± 2.1 37.9 ± 1.3 36.9 ± 2.5 Rye 35.8 ± 2.0 29.2 ± 1.9Chenopodium 33.9 ± 4.3 44.9 ± 17 38.7 ± 3.4 32.7 ± 2.4 Rape 44.3 ± 2.8 35.3 ± 2.4Summer mix 35.4 ± 1.3 34.9 ± 4.3 39.5 ± 1.1 34.7 ± 2.2 Winter mix 37.7 ± 2.0 29.0 ± 0.6

PLFAgram− nmol g−1 DM Oat 4.7 ± 0.3 4.3 ± 0.3 5.7 ± 0.3 4.6 ± 0.3 Hairy vetch 5.5 ± 0.6 5.4 ± 1.3Clover 5.0 ± 0.2 4.4 ± 0.3 5.1 ± 0.2 5.3 ± 0.4 Rye 4.7 ± 0.2 4.0 ± 0.3Chenopodium 4.9 ± 0.6 4.8 ± 0.7 5.3 ± 0.5 4.6 ± 0.3 Rape 5.8 ± 0.3 4.8 ± 0.3Summer mix 5.3 ± 0.1 5.2 ± 0.6 5.2 ± 0.1 4.7 ± 0.3 Winter mix 4.9 ± 0.3 3.9 ± 0.1

PLFAfungal nmol g−1 DM Oat 6.1 ± 1.0 5.7 ± 0.9 6.2 ± 0.7 5.0 ± 0.4 Hairy vetch 13.2 ± 1.6 7.8 ± 1.5Clover 8.7 ± 1.1 6.6 ± 0.9 6.0 ± 0.5 6.2 ± 0.6 Rye 10.2 ± 0.7 7.8 ± 0.4Chenopodium 6.6 ± 1.4 7.0 ± 1.2 6.0 ± 0.7 5.3 ± 0.6 Rape 12.3 ± 0.8 11.3 ± 1.4Summer mix 7.0 ± 0.4 7.5 ± 1.0 5.8 ± 0.3 6.0 ± 1.4 Winter mix 11.9 ± 1.5 8.5 ± 0.3

39K.A. Mackie et al. / Science of the Total Environment 500–501 (2014) 34–43

the vine than that at 120 cm in May 2013 as well as greater than bothdistances in June 2012.

3.4.3. Multivariate analysesThe multivariate analysis of the bacterial and fungal PLFAs indicated

that date (F = 45.98, p b 0.0001) and distance (F = 6.20, p b 0.0001)from the vine row were highly significant. To take a deeper look intohow the PLFAs separated, a PCA was performed (Fig. 3). The first twoprincipal components (PCs) together accounted for 81% of the PLFA var-iance for the entire year. May 2013 is clearly separated from August2012 and June 2012 by both PC1 and PC2. PC2 separates August andJune 2012. When the scores of the PCs were correlated with abioticsoil as well as plant properties, it can be seen that PC1 was significantlypositively correlated to EOC and plant biomass, and negatively correlat-ed to plant Cu content (Table 4). PC2 was significantly positively corre-lated to soil moisture. The loadings for each PC indicated that PC1 wasmainly associated with the PLFA fungal marker, while PC2 was mainlyassociated with gram-positive bacteria and eubacteria, which can beGram+ and Gram− (Table 4).

4. Discussion

4.1. Aboveground vegetation and copper uptake

4.1.1. Copper uptakeThis study is one of few that have sampled plants growing in situ for

phytoextraction purposes (Clemente et al., 2005, 2006; Brej andFabiszewski, 2006). Although there is a clear distinction between themaximum plant shoot Cu content, the amount of Cu accumulated perplant shoot DM, achievable by the treatments, the quantities suggestedin the literature were much higher than the quantities observed in thisstudy. Oat has been previously researched by Andreazza et al. (2010) ina greenhouse experimentwith two long-term vineyard soils (pH: 6–6.3,140–200 mg CuT kg−1 soil). The control treatment showed that oatshoots contained approximately 55 mg Cu kg−1 DM, nearly five timesthe amount measured in our field study. Reseda, although not showndue to its sporadic germination, was roughly five times lower than sug-gested by the literature (51 mg Cu kg−1 DM) (Poschenrieder et al.,2001). In 2010, Zeremski-Škorić et al. (2010) conducted a pot experi-ment of B. napus (pH: 7.2, 250 mg Cu kg−1 soil) and observed a shoot

Fig. 2.Microbial biomass (a), arylsulfatase (b) and phosphatase (c) during summer crops (June and August 2012) andwinter crop (May 2013) at a distance of 70 cm and 120 cm from thevine row. Where O is oat, CC is clover, CH is Chenopodium, SM is summer mix, HV is hairy vetch, RY is rye, RA is rape, and WM is winter mix. Mean ± SE is shown.

40 K.A. Mackie et al. / Science of the Total Environment 500–501 (2014) 34–43

content of 16.6 mg Cu kg−1 DM, approximately four times greater thanin our study. These comparisons demonstrate that field results are lessoptimistic than those observed in the laboratory. This could be becausethe Cu levels were higher in the lab experiments as well as suboptimalenvironmental factors in the field, such as less available water.

CuT and CuDTPA were distinguished by distance from the grapevines,where higher levels of Cu were nearest to the vine rows. Plant Cu

Fig. 3. Principal component analysis (PCA) of phospholipid fatty acids (PLFAs) in June2012, August 2012 and May 2013. Together PC1 and PC2 account for 81% of the variationamongst the PLFAs. The loadings (species scores) for the PLFAs are shown by the vectors.

content, however, was not significantly differentiated by distance. Itwas also, therefore, not affected by soil Cu content. Only treatmentplayed a significant role in determining how much Cu plants wereable to amass. This was also seen by Brej and Fabiszewski (2006),where in their field experiment only plant species, rather than heavymetal quality, played a role in plant Cu concentration.

4.1.2. Summer cropSummer crops significantly increased their plant Cu content from

June to August 2012, showing that the longer the investigated plantsgrew, the more Cu plants accumulated per DM. Despite this they didnot significantly increase their biomass over time, tending more to-wards a decrease in biomass with plant senescence. Moreover, highCu accumulating plants were not high biomass producers, as plant bio-masswas negatively correlated to plant Cu content. Therefore, althoughclover had the highest Cu content, followed closely by summermix, oathad the highest plant biomass at a distance of 120 cm. Biomass produc-tion significantly influenced Cu removal, the amount of Cu removedfrom the soil per area, more than plant shoot Cu content. This is becausebiomass and Cu removal are significantly positively correlated. At120 cm from the vine row, oat attained the most successful summerCu removal rate. Andreazza et al. (2010) estimated that oats could re-move approximately 175,000–200,000 mg Cu ha−1 for a normal vine-yard soil, assuming a biomass of 4000 kg DM ha−1. In our study, oat

Table 4Correlations of scores of principal component analysis for microbial communities (PLFAdata) with abiotic properties and plant characteristics for summer and winter cropscombined. Significant correlations (p b 0.05, r N 0.50) with axes PC1 and PC2 are shown.Properties in italics indicate negative correlations.

PLFAs Abiotic parameter

PC1 PC2 PC1 PC2

18.2ω6 i15.0 EOC Soil moisturea15.0 Plant biomass16.1ω7 Plant Cu content

41K.A. Mackie et al. / Science of the Total Environment 500–501 (2014) 34–43

produced 898 kg DM ha−1, which was much lower than that assumedby Andreazza et al. (2010), most likely due to variations in fertilization,plant competition and water availability. These lower biomass results,compared to the lab study, led to less Cu removal than estimated inthe literature. However, more biomass production during this periodof the year would have been impractical for the winegrower. This is be-cause growing oat and rye requires an increased input of water and fer-tilizer on the part of the winegrower. Moreover, the increased Cucontent over time did not have any significant effect on Cu removalover time. This indicated that the normal mowing date (June 2012)was sufficient for Cu extraction because there was not a significant in-crease in Cu removal during the two-month time extension.

4.1.3. Winter cropAlthough summer cropsweremore successful in accumulating Cu in

their plant shoots, winter crops were significantly better at producinghigher amounts of plant biomass. Rye was the most successful treat-ment compared to all others in June 2012 and May 2013. This was pos-sible despite an unusually long 2012/2013 winter, probably becausemore water was available in May 2013 as well as higher activity of mi-croorganisms in the spring when the crops began to grow vigorously.Therefore, the winter crops were significantly more successful in re-moving Cu. These results support the current practice, where covercrops are grown over winter due to more available water and less com-petition with grapevines (Steenwerth and Belina, 2008; Celette andGary, 2013).

4.1.4. Copper phytoextraction potentialDistance from the grapevine was significant for plant biomass,

which was higher at a distance of 120 cm compared to 70 cm. This isin contrast to the patterns displayed by CuT and CuDTPA, suggesting ei-ther that the plants grew better in lower Cu soil, which affected overallpotential for Cu removal, or that the plants avoided direct competitionwith the grapevines. As Cu was not negatively correlated to plant bio-mass, plant Cu content or Cu removal, it could be assumed that resourcecompetition, especially for sunlight, was themajor influencing factor. Asbiomass has a large overall effect on which plant species can be used toefficiently remove Cu from soil in situ, both species chosen, as well asseeding placement, have an influence on Cu phytoextraction (Kiddet al., 2009; Bhargava et al., 2012).When calculating the annual removalrate of summer plus winter crops, the best-case scenario is whenoat (approx. 15,000 mg Cu ha−1) and hairy vetch (approx.18,000 mg Cu ha−1) are based on growth 120 cm from the vine row;then there is the potential to remove 0.033 kg Cu ha−1 y−1. These re-sults indicate that phytoextractionwill not be a sufficientmethod for re-moving Cu in this vineyard soil. Initially, there was 95.9 kg Cu ha−1 inthe soil to a depth of 10 cm; in addition, 4 kg Cu ha−1 can be applied an-nually in Switzerland (Bio Suisse Standards, 2012). An extraction rate of0.033 kg Cu ha−1 y−1 will not balance the amount applied nor removea significant amount of Cu from the soil. It might be that whenadditionally removing the plant roots, as the literature suggests(Poschenrieder et al., 2001; Andreazza et al., 2010; Zeremski-Škorićet al., 2010), more Cu could be harvested. However, this would hardlybe feasible in practice.

4.2. Soil properties, microbial abundance and activity

4.2.1. Impact of plants on soil microorganismsThis study did not find any interaction between plants andmicrobial

properties. In contrast to the results aboveground, those belowgroundvaried only by distance from the vine row and date and showed no cor-relation to properties aboveground. There was also no beneficial impactfrom a diverse plant system as hypothesized. Kulmatiski and Beard(2011), who also did not see any influence of plants on the microbialcommunity in a short-term experiment, suggested that plants leave along-term microbial legacy in the soil, which may take years to adjust.

In a four year study in Germany, Habekost et al. (2008) also found littleevidence to show a change in soil microbial community due to plant di-versity, concluding that there is a significant time-lag belowgroundwhen aboveground modifications are made.

4.2.2. Impact of copper on soil microorganismsIn spite of the fact that high C and N pools coincided with high Cu

areas, indicated by their positive correlations and likely due to thebond between Cu and organic matter, microbial biomass and microbialfunction, i.e. enzyme activity, were either also high in these areas orwere not negatively affected. AlthoughWightwick et al. (2013) showedthat vineyards had higher Cu (approx. 100 mg Cu kg−1) and lower en-zyme activity (100–400 μg p-nitrophenol/phosphatase g−1 h−1) thanin reference soils, and Mackie et al. (2013) showed that Cu (approx.140 mg Cu kg−1) negatively affected phosphatase activity(813 μg phenol g−1 3 h−1), this study showed neither low values nor anegative correlation of enzymes to Cu quantities. This could be becausethe relative difference in Cu concentrationwas lower in this study or be-cause of different soil types, which affect pH, carbon and Cu speciation.

PLFAs were also not influenced by Cu in this study, suggesting thatsoil microorganisms are regulated by resource pools for survival andnot negatively by environments of pollution. Therefore, chemical prop-erties may be a more important factor for microorganisms than heavymetals are a deterrent, particularly in environments of low to moderatecontamination (Zhang et al., 2006; Wightwick et al., 2013). Zhang et al.(2006) suggest that this is due to overall resistance and resilience overtime. This is supported by Brandt et al. (2010), where the authorsshowed that entire bacterial communities could develop Cu tolerancewithout changing their structure and that these communities couldeven withstand a small increase in Cu pollution. It can be concludedthat transformations resulting from changes in Cu quantity and above-ground vegetation species occur over a period longer than one year.As there are neither vineyards available that are completely uncontam-inated nor vineyards that have significant Cu gradations within onearea, it is difficult to say at which Cu level toxicity begins occurring.Brun et al. (2003) suggested that the plant toxicity limit is hit whencontamination is greater than 250 mg CuT kg−1, while Mackie et al.(2013) saw microbial function impaired when contamination was140mg CuT kg−1. It can be conjectured that varying environmental fac-tors, plant presence, nutrient availability, and microbial communitiescould change this limit at different sites.

4.2.3. Seasonal variationsIt was observed that soil microorganism abundance changed over

the seasons. During the summer crop season, Cmic and Nmic were signif-icantly lower in abundance than those in May 2013. These observationscould be explained by a significant positive correlation between extract-able organic nutrients andmicrobial biomass and arylsulfatase. Aswaterand nutrients increased so did the microorganisms.

Seasonal shifts were also observed in the microbial community, asrepresented by PLFAs and differentiated by PCA. The fact that differentgroups of themicrobial communityweremore associatedwith differentenvironmental parameters, which coincided with crop season, suggeststhat therewas a change in dominance over the year (Regan et al., 2014).PLFAfungal had a strong association with EOC and plant biomass, indicat-ing an indirect reliance on nutrient pools most likely provided by rootexudates (Khalid et al., 2007). Habekost et al. (2008) also found thatPLFAs increased in abundance due to the quality and abundance of re-sources. Although there was no treatment specific response from themicrobial community, the presence of high plant biomass seemed to in-directly drive the abundance of the fungal community. Gram+bacteria,on the other hand, exhibited a strong relationship to soil moisture, im-plying that Gram+ bacteria are more influenced by physical driversthan by nutrient resources. This supports observations made by Reganet al. (2014) and Lennon et al. (2012), where decreasing environmentalstress factors, such as increases in soil moisture and extractable

42 K.A. Mackie et al. / Science of the Total Environment 500–501 (2014) 34–43

nutrients, increase fungal and bacterial abundance. Lennon et al. (2012)additionally observed that specific taxonomic groups responded differ-ently to changes in a moisture gradient, where Gram+ bacteria, in par-ticular Actinobacteria, were found to have a wetmoisture optimumanda narrow niche. Soil moisture and extractable nutrients supported theabundance of microorganisms, while likely soil moisture and microor-ganisms supported plant biomass production, all of which were attheir peak in May 2013. This is relevant because further investigationsin microbially-assisted phytoextraction will require the success of soilmicroorganisms and plants together and this has been shown to varyby season.

5. Conclusion

When considering overall Cu removal capabilities, the plants withthe most potential in a Wallisian vineyard were oat, hairy vetch andrye. However, the maximum removal rate, at 120 cm from the vinerow, was 0.033 kg Cu ha−1 DM y−1. These quantities were achievablebecause of the plants' high biomass production and not their shoot Cucontent.Winter cover crops, in particular, achieved the highest removalrates becausemorewaterwas available, more nutrients were accessibleand microorganisms were more active and abundant. Nevertheless,these in situ removal rates were too low to make phytoextraction ofCu in vineyards feasible, especially as Cu continues to be applied. Con-trary to our hypothesis that plant diversity would increase microbialabundance and diversity, mixed plots had no significant influence onsoil microbial properties.

Microbial abundance and activity were clearly regulated by Cand N pools despite the fact that these were also areas of high Cu.This suggests that organic matter and nutrient contents are maskingor reducing the negative effects of heavy metals. Two differentdrivers, soil moisture and extractable carbon, divided microbialcommunity structure into groups based on fungi and Gram+ bacteria.Abundance of the microbial community (Cmic, PLFAs) increased atthe same time as biomass production of plants, in May 2013. Futureperspectives of phytoextraction are focused on microbially assistedphytoextraction. Current literature shows that Cu phytoextractionby oat could be increased by approximately 200% with the inocula-tion of siderophore producing bacteria (Andreazza et al., 2010).This potential removal rate would remain well below the amountof Cu applied annually. In situ research is needed to further investigateplant species, microbial-plant interactions and microbial inoculationin soil.

Acknowledgments

We would like to thank Till Haas, Zorica Kauf, Matti Hanisch, RunaBoeddinghaus, Richard Ebner, Lisa Ebner, Felix Hegwein, and SilkeGrünewald for their assistance in the field as well as Claudio Niggli formanaging the field site and providing beneficial cover crop information.Thank you especially to Ibrahim Köran for his assistance in the field andin the lab and Juan Carlos Laso Bayas from the Department of Bioinfor-matics for his statistical consultations and assistance. We would alsolike to thank the Landesanstalt für Landwirtschaftliche Chemie andthe Soil Biology group at the University of Hohenheim for their support.Finally,wewould like to thank the Carl Zeiss Stiftung for funding this re-search and the anonymous reviewers for their assistance in improvingour paper.

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.scitotenv.2014.08.091.

References

Andreazza R, Okeke BC, Lambais MR, Bortolon L, de Melo GWB, de Oliveira Camargo FA.Bacterial stimulation of copper phytoaccumulation by bioaugmentation with rhizo-sphere bacteria. Chemosphere 2010;81:1149–54.

Andreazza R, Bortolon L, Pieniz S, Giacometti M, Roehrs DD, Lambais MR, et al. Potentialphytoextraction and phytostabilization of perennial peanut on copper-contaminatedvineyard soils and copper mining waste. Biol Trace Elem Res 2011;1–11.

Bhargava A, Carmona FF, Bhargava M, Srivastava S. Approaches for enhancedphytoextraction of heavy metals. J Environ Manage 2012;105:103–20.

Bio Suisse Standards. Downloaded on 04.11.2013 from http://www.bio-suisse.ch/en/li-brary/import/standards.php, 2012.

Brandt KK, Frandsen RJN, Holm PE, Nybroe O. Development of pollution-induced communitytolerance is linked to structural and functional resilience of a soil bacterial communityfollowing a five-year field exposure to copper. Soil Biol Biochem 2010;42:748–57.

Brej T, Fabiszewski J. Plants accumulating heavy metals in the Sudety Mts. Acta Soc BotPol 2006;75:61–8.

Brun LA, Maillet J, Richarte J, Herrmann P, Remy JC. Relationships between extractablecopper, soil properties and copper uptake by wild plants in vineyard soils. EnvironPollut 1998;102:151–61.

Brun LA, Le Corff J, Maillet J. Effects of elevated soil copper on phenology, growth andreproduction of five ruderal plant species. Environ Pollut 2003;122:361–8.

Castaldi P,Melis P, Silvetti M, Deiana P, Garau G. Influence of pea andwheat growth on Pb,Cd, and Zn mobility and soil biological status in a polluted amended soil. Geoderma2009;151:241–8.

Celette F, Gary C. Dynamics of water and nitrogen stress along the grapevine cycle as af-fected by cover cropping. Eur J Agron 2013;45:142–52.

Chen Y, Wang Y, Wu W, Lin Q, Xue S. Impacts of chelate-assisted phytoremediation onmicrobial community composition in the rhizosphere of a copper accumulator andnon-accumulator. Sci Total Environ 2006;356:247–55.

Clemente R, Walker DJ, Bernal MP. Uptake of heavy metals and As by Brassica junceagrown in a contaminated soil in Aznalcóllar (Spain): the effect of soil amendments.Environ Pollut 2005;138:46–58.

Clemente R, AlmelaC, BernalMP. A remediation strategybased on active phytoremediationfollowed by natural attenuation in a soil contaminated by pyrite waste. Environ Pollut2006;143:397–406.

DIN ISO 11466. Bodenbeschaffenheit — Extraktion in Königswasser löslicherSpurenelemente; 1995.

Epelde L, Becerril JM, Barrutia O, González-Oreja JA, Garbisu C. Interactions between plantand rhizosphere microbial communities in a metalliferous soil. Environ Pollut 2010;158:1576–83.

Epelde L, Becerril JM, Alkorta I, Garbisu C. Adaptive long-term monitoring of soil health inmetal phytostabilization: ecological attributes and ecosystem services based on soilmicrobial parameters. Int J Phytoremediation 2014;16:971–81.

Ettler V, Mihaljevič M, Šebek O, Grygar T. Assessment of single extractions for the deter-mination of mobile forms of metals in highly polluted soils and sediments—analyticaland thermodynamic approaches. Anal Chim Acta 2007;602:131–40.

European Commission. Council Regulation (EC). No 834/2007; 2007.Fernández-Calviño D, Soler-Rovira P, Polo A, Díaz-Raviña M, Arias-Estévez M, Plaza C. En-

zyme activities in vineyard soils long-term treated with copper-based fungicides. SoilBiol Biochem 2010;42:2119–27.

Frostegård A, Bååth E. The use of phospholipid fatty acid analysis to estimate bacterial andfungal biomass in soil. Biol Fertil Soils 1996;22:59–65.

Frostegard A, Tunlid A, Baath E. Microbial biomass measured as total lipid phosphate insoils of different organic content. J Microbiol Methods 1991;14:151–63.

Frostegård Å, Tunlid A, Bååth E. Changes in microbial community structure during long-term incubation in two soils experimentally contaminated with metals. Soil BiolBiochem 1996;28:55–63.

Ge C, Zhang Q. Microbial community structure and enzyme activities in a sequence ofcopper-polluted soils. Pedosphere 2011;21:164–9.

Gómez-Sagasti MT, Alkorta I, Becerril JM, Epelde L, AnzaM, Garbisu C. Microbial monitoringof the recovery of soil quality during heavy metal phytoremediation. Water AirSoil Pollut 2012;223:3249–62.

Guerra B, Steenwerth K. Influence of floor management technique on grapevine growth,disease pressure, and juice and wine composition: a review. Am J Enol Vitic 2012;63:143–64.

Habekost M, Eisenhauer N, Scheu S, Steinbeiss S, Weigelt A, Gleixner G. Seasonal changesin the soil microbial community in a grassland plant diversity gradient four yearsafter establishment. Soil Biol Biochem 2008;40:2588–95.

Haferburg G, Kothe E. Metallomics: lessons for metalliferous soil remediation. AppMicrobiol Biotechnol 2010;87:1271–80.

Heibertshausen D, Hofmann U, Baus-Reichel O, Berkelmann-Loehnertz B. Copper replacementand copper reduction in organic viticulture by the use of biopesticides andnew copper for-mulations. Bulletin OILB/SROP 2006;29:23–6.

Hinojosa M, Garcia-Ruiz R, Carreira J, Karaca A, Cetin S, Turgay O. Soil biology: soil heavymetals. Berlin: Springer; 2010.

Joergensen RG. The fumigation–extraction method to estimate soil microbial biomass:calibration of the kEC value. Soil Biol Biochem 1996;28:25–31.

Kandeler E, Kampichler C, Horak O. Influence of heavy metals on the functional diversityof soil microbial communities. Biol Fertil Soils 1996;23:299–306.

Kandeler E, Mosier AR, Morgan JA, Milchunas DG, King JY, Rudolph S, et al. Transientelevation of carbon dioxide modifies the microbial community composition in asemi-arid grassland. Soil Biol Biochem 2008;40:162–71.

Khalid M, Soleman N, Jones DL. Grassland plants affect dissolved organic carbon andnitrogen dynamics in soil. Soil Biol Biochem 2007;39:378–81.

43K.A. Mackie et al. / Science of the Total Environment 500–501 (2014) 34–43

Kidd P, Barceló J, Bernal MP, Navari-Izzo F, Poschenrieder C, Shilev S, et al. Trace elementbehaviour at the root–soil interface: implications in phytoremediation. Environ ExpBot 2009;67:243–59.

Kos B, Leštan D. Chelator induced phytoextraction and in situ soil washing of Cu. EnvironPollut 2004;132:333–9.

Kulmatiski A, Beard KH. Long-term plant growth legacies overwhelm short-term plantgrowth effects on soil microbial community structure. Soil Biol Biochem 2011;43:823–30.

La Torre A, Spera G, GianferroM, ScaglioneM.More years of field trials against Plasmoparaviticola in organic viticulture. Commun Agric Appl Biol Sci 2007;72:901–8.

Lennon JT, Aanderud ZT, Lehmkuhl BK, Schoolmaster Jr DR. Mapping the niche space ofsoil microorganisms using taxonomy and traits. Ecology 2012;93:1867–79.

Ma Y, Rajkumar M, Freitas H. Inoculation of plant growth promoting bacteriumAchromobacter xylosoxidans strain Ax10 for the improvement of copperphytoextraction by Brassica juncea. J Environ Manage 2009;90:831–7.

Mackie KA, Müller T, Kandeler E. Remediation of copper in vineyards— amini review. En-viron Pollut 2012;167:16–26.

Mackie KA, Müller T, Zikeli S, Kandeler E. Long-term copper application in an organicvineyard modifies spatial distribution of soil micro-organisms. Soil Biol Biochem2013;65:245–53.

McBride M, Tiller K, Merry R. Copper in soils and plants. Sydney: Academic Press; 1981.Meier S, Borie F, Bolan N, Cornejo P. Phytoremediation of metal-polluted soils by

arbuscular mycorrhizal fungi. Crit Rev Env Sci Technol 2012a;42:741–75.Meier S, Borie F, Curaqueo G, Bolan N, Cornejo P. Effects of arbuscular mycorrhizal inocu-

lation on metallophyte and agricultural plants growing at increasing copper levels.Appl Soil Ecol 2012b;61:280–7.

Moolenaar SW. Sustainable management of heavy metals in agro-ecosystems.Wageningen, Netherlands: Landbouwuniversiteit Wageningen; 1998.

Morlat R, Jacquet A. Grapevine root system and soil characteristics in a vineyard main-tained long-term with or without interrow sward. Am J Enol Vitic 2003;54:1–7.

Mueller T, Joergensen RG, Meyer B. Estimation of soil microbial biomass C in the presenceof living roots by fumigation–extraction. Soil Biol Biochem 1992;24:179–81.

Narula N, Kothe E, Behl RK. Role of root exudates in plant–microbe interactions. J Appl BotFood Qual 2009;82:122–30.

Paoletti MG, Sommaggio D, Favretto MR, Petruzzelli G, Pezzarossa B, Barbafieri M. Earth-worms as useful bioindicators of agroecosystem sustainability in orchards andvineyards with different inputs. Appl Soil Ecol 1998;10:137–50.

Poschenrieder C, Bech J, LluganyM, PaceA, Fenés E, Barceló J. Copper in plant species in a cop-per gradient in Catalonia (North East Spain) and their potential for phytoremediation.Plant Soil 2001;230:247–56.

R Core Team. R: a language and environment for statistical computing. 0.97.551; 2013.Rajkumar M, Ae N, Prasad MN, Freitas H. Potential of siderophore-producing bacteria for

improving heavy metal phytoextraction. Trends Biotechnol 2010;28:142–9.Regan KM, Nunan N, Boeddinghaus RS, Baumgartner V, Berner D, Boch S, et al. Seasonal

controls on grassland microbial biogeography: are they governed by plants, abioticproperties or both? Soil Biol Biochem 2014;71:21–30.

Salinari F, Giosuè S, Tubiello FN, Rettori A, Rossi V, Spanna F, et al. Downy mildew(Plasmopara viticola) epidemics on grapevine under climate change. Glob ChangBiol 2006;12:1299–307.

Schinner F, Oehlinger R, Kandeler E, Margesin R. Methods in soil biology. Berlin: SpringerVerlag; 1996.

Song J, Zhao F, Luo Y, McGrath SP, Zhang H. Copper uptake by Elsholtzia splendens andSilene vulgaris and assessment of copper phytoavailability in contaminated soils. En-viron Pollut 2004;128:307–15.

Steenwerth K, Belina KM. Cover crops and cultivation: impacts on soil N dynamics andmicrobiological function in a Mediterranean vineyard agroecosystem. Appl Soil Ecol2008;40:370–80.

Terry N, Bañuelos G. Phytoremediation of contaminated soil and water; 2000.Vance ED, Brookes PC, Jenkinson DS. An extraction method for measuring soil microbial

biomass C. Soil Biol Biochem 1987;19:703–7.VDLUFA. Methodenbuch Bd. VII, Methode 2.1.1: Nassaufschluss unter Druck. Speyer,

Germany: VDLUFA-Verlag; 2011a.VDLUFA.MethodenbuchBd.VII,Methode2.2.2.6: BestimmungvonausgewähltenElementen

in pflanzlichemMaterial und Futtermitteln mit Optische Emissionsspektrometrie mitinduktiv gekoppeltem Plasma (ICP-OES). Speyer, Germany: VDLUFA-Verlag; 2011b.

VDLUFA. Methodenbuch Bd. I, Methode A 6.4.1: Bestimmung von Mg, Na und denSpurenelementen Cu, Mn, Zn und B im Calciumchlorid/DTPA-Auszug (CAT). Speyer,Germany: VDLUFA-Verlag; 2011c.

Wardle DA, Bardgett RD, Klironomos JN, Setälä H, Van Der Putten WH, Wall DH. Ecolog-ical linkages between aboveground and belowground biota. Science 2004;304:1629–33.

Wenzel WW. Rhizosphere processes and management in plant-assisted bioremediation(phytoremediation) of soils. Plant Soil 2009;321:385–408.

Wightwick AM, Salzman SA, Reichman SM, Allinson G, Menzies NW. Effects of copperfungicide residues on the microbial function of vineyard soils. Environ Sci PollutRes 2013;20:1574–85.

Yang R, Tang J, Chen X, Hu S. Effects of coexisting plant species on soil microbes and soilenzymes in metal lead contaminated soils. Appl Soil Ecol 2007;37:240–6.

Zelles L. Fatty acid patterns of phospholipids and lipopolysaccharides in the charac-terisation of microbial communities in soil: a review. Biol Fertil Soils 1999;29:111–29.

Zeremski-Škorić TM, Sekulić PD, Maksimović IV, Šeremešić SI, Ninkov JM, Milić SB, et al.Chelate-assisted phytoextraction: Effect of EDTA and EDDS on copper uptake by Bras-sica napus L. J Serb Chem Soc 2010;75:1279–89.

Zhang C, Huang L, Luan T, Jin J, Lan C. Structure and function of microbial communitiesduring the early stages of revegetation of barren soils in the vicinity of a Pb/Zn smelt-er. Geoderma 2006;136:555–65.